/predictionary

A learning JavaScript dictionary-based word prediction / autocomplete / suggestion library.

Primary LanguageJavaScriptGNU Affero General Public License v3.0AGPL-3.0

Predictionary

JavaScript dictionary-based word prediction library with self-learning abilities.

Installation

Drop the following into your page:

<script src="https://unpkg.com/predictionary/dist/predictionary.min.js"></script>

or install the package via npm:

npm install predictionary --save

Note: to use inside a nodejs application use the flag --experimental-modules in order to be able to use Predictionary which is written as ES6 module. See nodejs demo which can be run with:

node --experimental-modules demo/node-demo/app.mjs

Basic usage

Minimum working example for basic usage:

import Predictionary from 'predictionary' //only if installed via npm

let predictionary = Predictionary.instance();
predictionary.addWords(['apple', 'apricot', 'banana']);
let suggestions = predictionary.predict('ap'); // == ['apple', 'apricot'];

Demo

See working demo: to live demo

API Documentation

see full API documentation: to API documentation

Acknowledgements

Thanks to Mark Davies, Professor of Linguistics at Brigham Young University (Utah), for the permission to use his n-gram sample data from https://www.ngrams.info/iweb.asp for training purposes of this library.

Word frequency lists for demos are taken from the Centre for Translation Studies, University of Leeds, see http://corpus.leeds.ac.uk/list.html - thanks!